Antibiotic use, hospital admissions, and mortality before and after implementing guidelines for nursing home-acquired pneumonia.

OBJECTIVES: To compare two strategies for implementing guidelines for nursing home-acquired pneumonia (NHAP) and to measure outcomes associated with treatment in accordance with the guidelines.
DESIGN: Randomized controlled trial.
SETTING: Ten skilled nursing facilities (SNFs) from a single metropolitan area. PARTICIPANTS: Patients with an episode of pneumonia acquired more than 3 days after admission to SNF (N = 350): 226 preintervention episodes of pneumonia and 116 postintervention episodes.
INTERVENTIONS: Multi-faceted education intervention including small-group consensus process limited to physicians and a similar intervention that included physicians and nurses within randomly selected SNFs. MEASUREMENTS: Antibiotic use at diagnosis compared with the guidelines, hospital admission, severity of pneumonia, and 30-day mortality.
RESULTS: Data were complete for 344 episodes of NHAP. For the preintervention group (n = 226), 62.2% (79/127) of the episodes were treated with parenteral antibiotics (PA) when PA were recommended by the guidelines and 57.6% (57/99) of episodes were treated with oral antibiotics (OA) when OA were indicated by the guidelines. Postintervention, treatment with PA and OA according to the guidelines was not significantly different between the two groups of randomized SNFs. A multivariate analysis comparing PA use pre- and postintervention for all SNFs, adjusted for variation in the frequency and severity of pneumonia, found significantly more of the postintervention episodes were treated with PA in accordance with the guidelines (P < .02). A preintervention significant difference in 30-day mortality observed between episodes with indications for PA (37.8% (48/127)) and episodes with indications for OA (6.1% (6/99)) (P < .001) was not present postintervention (11.5% (6/52); (23.8% (15/64); P = .06). There was no significant difference in 30-day mortality preintervention and postintervention for episodes with guideline indications for OA (P = .35) or for PA (P = .05) (P = .16 for multivariate analysis). The difference in PA use was not associated with significant differences in hospital admissions for episodes on NHAP.
CONCLUSION: The increase in the use of PA provides evidence that care within SNFs can be significantly changed using standard quality improvement techniques. Use of the guidelines did not significantly affect mortality. The addition of a practical severity of NHAP model or a change in reimbursement structure may enhance the guidelines' impact on hospitalization for NHAP. The financial benefits available with use of the guidelines will be limited unless the guidelines contribute to a reduction in rates of hospitalization.

RCT Entities:   Population Interventions Outcomes